Clip: In which Sherlock pops his collar.
It’s a bold move, questioning Sherlock Homes. But the famous investigator just begs for scrutiny in the second episode of the BBC’s hit series Sherlock. In the spirit of Dr. John Watson, I’ll enjoy setting him right.
Sherlock makes a precise claim relating to cognitive science — specifically visual memory — during the show’s second episode, “The Blink Banker.” It’s nighttime, and he and John are looking for graffiti that John spotted earlier near some railway tracks. But they discover that the graffiti has since disappeared. So Sherlock grabs John by the shoulders, spins him around and commands him to close his eyes to “maximize” his visual memory. Remember, Sherlock notes, “The average human memory on visual matters is only 62% accurate.”
Wait, 62% accurate? That claim made my ears perk up. The number is so specific, and in over four years of studying human cognition, I have never heard that statistic. It continued to elude me after lots of internet searching (clearly the gold standard for research). So in a fit of Sherlock-esque snooping, I turned to a friend who knows more about memory than I do. She tracked down one 62% figure in relation to visual working memory, in a paper by researchers at Bangor University.
As the paper explains, researchers got that figure by studying people’s performance in a visual search task. The study participants stared at a bunch of weird shapes and looked for whichever shape the researchers had asked them to find. While doing this, they were also supposed to remember (i.e., continuously picture in their minds) a different shape, which they would later be asked to report. Participants were able to do both things — find the first shape and remember the second — 62% of the time on average.
I don’t know if this is the statistic Sherlock was referencing. John wouldn’t necessarily have used working memory to find the graffiti, since working memory usually refers to memory storage over a very short period of time. But even if Sherlock specifically meant visual working memory, he would be making an overstatement.
That the people in this study could, about two-thirds of the time, hold one image in memory while also locating another shape points to cool possible insights about visual multitasking. But of course, it does not mean that human memory on visual matters is generally 62% accurate. Looking for funny shapes in a quiet, controlled lab differs from many every day visual memory experiences, such as — hypothetically — trying to remember in which aisle the local grocery store stocks Cadbury cream eggs while pretending to pay attention in class. So unless Sherlock’s writers can turn up a more convincing paper, I throw down my gloves and say, prove it!
I will also note, I’m onto Sherlock’s game of saying things really confidently so people just take for granted that he is right. That kind of speech is a very effective way to show off or bully others during arguments. Like many clever people, Sherlock could probably do with a reminder that he’s not as clever as he thinks he is. Unless becoming more modest would stop him popping his collar — that just wouldn’t be worth it.